388 Mr. H. C. Jones on the Cryoscopic Relations 



agree to within experimental errors, and are in all cases 

 much greater than the theoretical value. 



These results do not differ to any considerable degree from 

 those which were found when cane-sugar was first studied*. 



It was then decided to make some determinations in which 

 the temperature of the freezing-mixture differed from the 

 freezing-point of the solution under investigation by an 

 amount so small that the latter only just froze. The 

 freezing-mixture used was prepared from alum and ice. 

 But on account of the large volume of the solution used 

 (1100 cubic centim.) and the large air-space which sepa- 

 rated this from the freezing-mixture, it was found that the 

 time required to cool the solution even o, 5 was so great as 

 to be impracticable. A freezing-mixture which answered 

 the requirements satisfactorily was prepared as follows : — An 

 intimate mixture was made of powdered ice and alum, then a 

 very small amount of common salt was added. This mixture, 

 when used in my apparatus, gave a temperature in the air- 

 bath around the solution of never more than 0°*5, and usually 

 less than 0°*4, lower than the freezing-point of the solution. 

 The solution was cooled down to within a fraction of a degree 

 from its freezing-point in a colder bath, and was then placed 

 in the bath surrounded by the mixture just described. All 

 determinations were thus made with a freezing-mixture from 

 o, 3 to 0°'4z colder than the freezing-point of the solution. 



The details of the procedure for a single determination 

 were as follows : — 1100 cubic centim. of distilled water, cooled 

 nearly to zero, were placed in the glass vessel Cf, and cooled 

 to within a tenth of a degree of zero in a bath surrounded 

 by a mixture of salt and ice. The glass vessel was then 

 placed in the metallic vessel B, which was surrounded by a 

 freezing-mixture of ice and alum prepared as above described, 

 and the water stirred gently until it had cooled down to — o, 6 

 or — 0°*7. A small piece of pure ice was then thrown into the 

 water, which brought about at once a separation of ice 

 throughout the liquid. The stirring during the formation of 

 the ice was accomplished by a very gentle movement of the 

 stirrer, which with the form employed was found to be amply 

 sufficient. In this manner the liberation of heat in such quan- 

 tities that the thermometer was affected by it before it could 

 meet the ice was prevented, as was shown by the fact that when 

 the thermometer had reached a constant point, stirring for a 

 long time failed to change the reading a perceptible amount. 

 Thus was avoided experimentally a comparatively large and 



* Zeit.f. jjhys.Chcm. xii. p. 642; and Phil. Mag. xxxvi. p. 486 (1893). 

 j Zeit.f. phys, ('hem. xi. p. 532. 



